On this page
A conductive adhesive attachment can retain electrical continuity while its position or reaction force changes under a sustained mechanical condition. Before comparing two cured joints, identify what the surrounding assembly holds constant. A hanging load and a rigid restraint ask different questions, even when both begin at the same force and displacement.
Key design decisions
- Define the controlled mechanical quantity at the actual joint, not only at the test-machine crosshead.
- Compare creep and relaxation using their own time histories rather than exchanging one fitted time constant.
- Retain electrical measurements as a separate functional response under the stated mechanical condition.
1. Define the movement that matters to the circuit
Choose a physical relative displacement between the attached terminal or component and a stated point on the ceramic. Movement at a remote lead tip includes lead bending and may not equal motion across the bondline. Draw the force direction, restraint locations and measurement endpoints before requesting a time-dependent adhesive property.
The relevant output may be terminal alignment, clearance to an adjacent conductor or force transferred into a fragile pad. These outputs are not interchangeable. A joint that moves acceptably under a suspended lead can still impose an excessive reaction on a rigid package. Keep the cured adhesive, terminal, pad and ceramic in the mechanical definition; do not substitute an uncured paste viscosity for a cured attachment response.
2. Separate a sustained force from an imposed displacement
In a creep experiment, the force on the defined assembly coordinate is held while displacement is observed. In a relaxation experiment, displacement is held while the reaction force is observed. For a homogeneous specimen these correspond to the familiar stress and strain descriptions, but an attachment often has a nonuniform local stress field.
A load frame holding its crosshead does not necessarily hold the joint displacement. Grip compliance, terminal bending, substrate motion and thermal expansion can all contribute. Record force and the local displacement together, even when only one is controlled. If both quantities vary appreciably, interpret the measured boundary condition rather than labelling the trace an ideal constant-load or fixed-displacement test.
3. Use a three-element model as a falsifiable approximation
Consider a single effective displacement coordinate with an equilibrium spring in parallel with a branch containing another spring and a dashpot in series. The equilibrium spring retains a force after the delayed branch has relaxed. This standard-linear-solid approximation distinguishes an immediate response from a delayed one without assuming that a real filled adhesive has only one physical relaxation mechanism.
Use effective attachment stiffnesses in newtons per millimetre and a dashpot coefficient in newton-seconds per millimetre. These are geometry-dependent parameters, not bulk material moduli. Require a stable cured state, controlled temperature, small reversible motion and no damage during the comparison. If the observed trace requires several distinct rates, a single exponential should not be forced through it merely because three numbers are convenient to report.
4. Calculate force decay for an imposed movement
Assume an equilibrium stiffness of 2 newtons per millimetre, a delayed-branch spring stiffness of 8 newtons per millimetre and a dashpot coefficient of 80 newton-seconds per millimetre. These deliberately assumed values describe a mathematical attachment, not a named adhesive. A rapid displacement step of 0.10 millimetre initially loads both springs.
The initial force is 1.00 newton and the long-time force is 0.20 newton. The relaxation time is 10 seconds. At 10 seconds, the predicted force is 0.20 plus 0.80 divided by e, approximately 0.4943 newton. A declining force in this calculation is not evidence of debonding: the intact model itself relaxes. Conversely, fitting the trace does not prove that an actual joint remained intact.
F(t) = δ0[k∞ + km exp(−t/τr)]; τr = η/km
- F is reaction force in N; δ0 is the held displacement in mm.
- k∞ and km are positive effective stiffnesses in N/mm; η is in N·s/mm.
- τr and t are in seconds; the initial effective stiffness is k0 = k∞ + km.
Initially unloaded linear standard-linear-solid coordinate, ideal displacement step and fixed temperature. Fixture compliance, damage and changing material state are excluded.
5. Do not reuse the relaxation time for creep
Now apply a force step of 1.00 newton to the same assumed model and hold that force. Its initial displacement is again 0.10 millimetre, but the long-time displacement becomes 0.50 millimetre. The creep retardation time is 50 seconds, five times the relaxation time, because the equilibrium stiffness is only one fifth of the initial stiffness.
At 50 seconds, displacement is 0.50 multiplied by the quantity one minus 0.80 divided by e, approximately 0.3528 millimetre. It would be incorrect to use the 10-second force-decay constant in this creep expression. The equal initial point does not make the later experiments equivalent. Plot the controlled input with the response so that a fit cannot conceal which experiment supplied the parameters.
δ(t) = (F0/k∞)[1 − (km/k0)exp(−t/τc)]; τc = ηk0/(k∞km) = τr k0/k∞
- F0 is the held force in N; δ is displacement in mm.
- k0 = k∞ + km and all parameters have the same definitions as the relaxation calculation.
- τc is the creep retardation time in seconds, not τr.
Same initially unloaded linear model and fixed material state, now with an ideal force step. This relation is not a fatigue, rupture-time or adhesive-strength prediction.
6. Challenge the model before extrapolating the record
Repeat the experiment at suitably chosen smaller amplitudes while keeping geometry, temperature and initial history comparable. In a linear regime, response normalized by the imposed amplitude should agree within measurement uncertainty. A departure can reflect nonlinear behavior, damage or an uncontrolled boundary; inspect those possibilities before fitting another stiffness.
Retain the actual loading rise time and earliest usable observation. A slow force ramp can hide the immediate displacement and bias the estimated initial stiffness. Likewise, a short observation window may not establish a long-time plateau. Report which parameters are constrained by measured data and which remain poorly determined. Do not extrapolate a short room-temperature trace to years of service or another temperature without a separate, validated aging and temperature model.
7. Assign separate decisions to force, movement and resistance
Mechanical and electrical records should share specimen identity and timestamps, but neither substitutes for the other. Preserve events during loading and after release so recovery can be distinguished from a retained change.
| Observation | Resolve first | Appropriate decision |
|---|---|---|
| Force falls under locally fixed displacement | Restraint stability, temperature and physical joint condition | Compare retained reaction with the specified load-transfer requirement |
| Movement grows under held force | Local motion versus lead or fixture motion | Compare the actual clearance or alignment coordinate |
| Resistance changes with little visible motion | Electrical boundary, temperature and conducting interfaces | Investigate the current path separately from the mechanical fit |
| Normalized traces change with amplitude | Linearity, damage and preparation consistency | Use an amplitude-specific response or a justified nonlinear model |
| Apparent plateau lies beyond the measured interval | Parameter identifiability and observation duration | Extend the record instead of assigning an unsupported equilibrium value |
8. Translate the result into an attachment requirement
Write the required force or displacement history into the assembly review, including dwell, unload sequence and temperature. Specify the quantity that limits function: for example, terminal position at a defined time or reaction force after a fixed mounting displacement. Neither quantity is automatically a universal shear-strength limit.
Maintain the actual bondline and terminal configuration when comparing alternatives. A changed thickness, overlap or lead compliance changes the effective mechanical coordinate even when the adhesive name is unchanged. For a mixed-compliance package, combine the attachment model with the surrounding structure or test that assembly directly. Keep acceptance tied to the measured configuration and retain separate electrical, physical-integrity and movement criteria.
Review a sustained-load conductive attachment
Provide the joint geometry and the mechanical history that the installed assembly imposes.
- Cured adhesive identity and history, bonded surfaces, bondline dimensions and terminal or component geometry.
- Force direction, restraint stiffness, local displacement endpoints, dwell and unload sequence.
- Temperature record, loading rise time, force and displacement traces, specimen preparation and recovery observations.
- Electrical measurement points and timestamps, allowable movement or retained force, and physical failure observations.
The drawing-upload form loads as you reach this section.